584 C. CLIFFORD DOBELL. 
the spore rudiment—stains deep red, like chromatin, with 
Giemsa’s stain. But a little later it changes its reaction, and 
is stained a bright blue (fig. 14). This, as I have already 
shown in B. flexilis, is owing to the formation of the spore 
membrane. As the membrane hardens the spore gradually 
stains less and less—until, finally, it refuses to stain at all 
(figs. 15, 16). At this stage the spore is fully formed, and 
the cell breaks up and liberates it more or less completely 
(fig. 16). It must be noted that a part of the filament per- 
sists outside the spore until quite a late stage, and breaks up 
with the rest of the cell (see figs. 14, 15, 16). 
I occasionally found degenerating bacilh occurring together 
with those which appeared quite normal. Some of the forms 
encountered are shown in figs. 17, 18, and 19, and will require 
no further description. 
A most remarkable—and, as I believe, important—variation 
in the method of forming spores was seen on a few occasions. 
The large original organism had failed to divide completely, 
and each of the small daughter individuals had developed 
spore rudiments whilst still attached to one another 
(fig. 13). [A discussion of these abnormal forms and their 
significance will be found on p. 587, et seq. | 
THe Merruop or SPorRE-FORMATION IN BACTERIUM LUNULA, 
N. SP. 
Under the name Bacterium lunula I will here describe 
another micro-organism which [| have found in the rectum of 
toads. It is, like the preceding form, very uncommon and 
of considerable size. It reaches a length of about 15 p, 
though an average size is about 10 .—12 p. 
The chief characteristic of this organism is its curved 
shape (fig. 20). The ordinary individuals appear to have a 
chromatic filament similar to that of B. spirogyra (cf. fig. 
20), but not so clearly defined. I cannot state definitely that 
